Thought Leadership

Before AI Savings Go into Your 2027 Claims Budget, Ask These Three Questions

By SiftMed | November 17, 2026

This Thought Leadership is part of a series brought to you by SiftMed.

View series

What to check before a productivity assumption becomes your team’s workload target.

Claims organizations are going into 2027 with big expectations on how AI will improve productivity.

Those expectations extend beyond the technology budget. They can influence individual workloads, service commitments, and the number of claims your organization expects each adjuster to manage.

A productivity gain on paper eventually becomes an expectation for the people doing the work. By January, that target could land on a senior adjuster’s desk alongside 140 open files, two inherited from a colleague who retired in October and a 2,600-page medical record that arrived late Friday.

Before building those gains into next year’s operating plan, claims leaders need to understand what work the technology will actually do, how much time it will save across the process, and what the organization can realistically do with the capacity it recovers.

Whether your plan is still under review or already signed off, now is the time to ask whether those expectations are realistic and what needs to change to make it possible.

1. Which parts of claims review are taking up your team’s time?

The number of claims an adjuster handles doesn’t tell you how much work each claim required. A lost-time comp claim with a decade of prior treatment involves a very different level of review than a soft-tissue claim with three clinic visits. An inherited file might take hours of work just to get caught up on the treatment history, outstanding questions, and decisions already made.

Medical record review involves more than reading clinical notes and reports. Adjusters may need to locate a specific detail in an IME report buried under 1,800 pages, identify duplicates, and re-order treatment dates as new records arrive.

These tasks can consume hours that could otherwise go toward assessing medical evidence and applying professional expertise. AI can help organize records, identify duplicates, build timelines or summaries and make information easier to find. The reviewer still needs to assess the evidence, consider causation and determine what matters to the claim.

2. How will claims teams measure whether AI is delivering the improvements you expect?

A good demo or impressive processing time doesn’t guarantee consistent results across every claim. A good demo earns a place in the planning conversation, not a number in the capacity model.

Start by establishing a baseline. How long does a medical record review take today? How much time do adjusters spend verifying information, correcting errors, or searching for details? What needs to be included before a claim is ready for the next step in the process?

Then decide what you’ll measure once AI is introduced. Processing time is one metric, but look at the time it takes to produce a result an adjuster can actually use and sign off on, not just a first draft.

Can reviewers trace findings directly to the original record and read them in context? Are summaries and timelines capturing the details adjusters need? If findings can’t be easily verified, the time hasn’t been saved. It has just moved.

See article on: 5 Reasons AI Rollouts Stall in Claims

Test those assumptions against your own claims before settling on a productivity target. Include a range of files, including duplicates, handwritten notes, and the 3,000-page litigated claim.

Account for training, rollout and oversight, too. These can all affect how much time your team actually gets back, so test across different claim types and levels of complexity before setting your productivity assumptions.

3. What will your organization do with the capacity it gains?

Hours saved are most valuable if you have a plan for what happens next. Before setting productivity targets, consider where additional capacity could make the biggest difference to your claims operation.

For some organizations, it might mean absorbing more claims without increasing caseload pressure. For others, it could mean giving experienced adjusters more time to focus on complex or litigated files, reducing overtime, improving cycle times, or creating opportunities to coach newer adjusters.

Integral National, an IME and peer review firm working no-fault and workers’ comp files, offers an example of what recovered time can look like. Client Coordination Supervisor Jennise Ortiz, who has 26 years in this space, explains that a file of 500-plus pages takes time to sort, and every change used to mean sending the file back and waiting a day or two. After SiftMed helped reduce their turnaround time, Jennise moved the recovered time to the Retrieval Team, which works directly with providers to obtain IME reports, peer reviews and addendums.

The important part is deciding where that time should go before it gets absorbed into existing workloads.

Be specific about the outcomes you’re planning for, and measure them accordingly. If the goal is to handle more claims, look at caseloads and cycle times. If it’s to give adjusters more time for complex work, look at how that time is being allocated.

Give the assumption an owner

Before January, bring claims departments, finance, security, and the people who will use the technology in one room. Talk about how and where you’ll leverage AI, agree on the task AI will support, the results you’ll use to measure its impact, and how the organization intends to use the capacity it recovered.

Assign someone to monitor results, and schedule a review about 90 days after implementation, then again 6 months down the road. Compare actual performance against the assumptions from your 2027 plan, adjust accordingly.

It’s better to revisit targets in April than to spend the rest of the year asking your team to meet a target that wasn’t realistic.

Behind many of those budget lines is an earlier question: build AI capability in-house or buy it? SiftMed’s CEO, Holly Hill, and Chief AI Officer, Amir Tavafi, walked through the trade-offs in a recent webinar. Get in touch with our team for the recording or to see how SiftMed handles one of your medical files.

__________________________

About SiftMed:

At SiftMed, we’re changing how professionals work with complex medical records. Our purpose-built AI platform brings order to thousands of pages of information, organizing documents, building summaries and timelines, and making it easy to search, explore and trace every finding back to its source.

We believe technology should make expertise more powerful, not replace it. That’s why we focus on making medical facts easier to find, understand and verify, so the people making important decisions have the information they need to stand behind them with confidence.

Was this article valuable?

Here are more articles you may enjoy.